You can deploy this model and start earning money today!

Jnx03/
kanitakorn-r1d-qwen14b-stage1-v1-exp-20260521

Name: kanitakorn-r1d-qwen14b-stage1-v1-exp-20260521
Availability: InStock

deepseek-ai/DeepSeek-R1-Distill-Qwen-14B

adapter

mit

Views

No views yet

Jnx03/kanitakorn-r1d-qwen14b-stage1-v1-exp-20260521

LoRA adapter from the Kanitakorn 7-day Thai LLM campaign.

Base: deepseek-ai/DeepSeek-R1-Distill-Qwen-14B
Stage: stage1-v1-exp
Date: 2026-05-21
Dataset SHA256: —

Hyperparameters

json

{
  "lr": 1e-05,
  "epochs": 1,
  "r": 16,
  "qlora": true,
  "mix": "30:70",
  "bs": 4,
  "ga": 2,
  "max_len": 1536
}

Evaluation (top-3 benchmarks)

Benchmark	Score	Target	Beat?
—	—	—	—

Recipe summary

LoRA rank: 16, alpha: —, dropout: —
Learning rate: 1e-05, epochs: 1
QLoRA 4-bit (nf4): True
Mix Thai:EN: 30:70

Usage

python

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "deepseek-ai/DeepSeek-R1-Distill-Qwen-14B"
adapter = "Jnx03/kanitakorn-r1d-qwen14b-stage1-v1-exp-20260521"

tok = AutoTokenizer.from_pretrained(base, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(base, torch_dtype="bfloat16", device_map="auto", trust_remote_code=True)
model = PeftModel.from_pretrained(model, adapter)

Campaign log

See VERSION_LOG.md for context on this checkpoint vs siblings.

License: MIT (inherited from deepseek-ai/DeepSeek-R1-Distill-Qwen-14B).